Due to recent advances in Magnet Resonance Imaging (MRI), there is a growing interest in using MRI to image the heart. Because the MRI image is very sensitive to motion from slice to slice, and stopping the heart is impractical, it is necessary to accurately detect the peak of the “R” wave of the ECG signal to generate a trigger signal and thereby insure that each image slice is taken when the heart is in the same relative position.
The traditional approach has been to use non-metallic electrodes and patient leads to bring the ECG signal out of the bore of the magnet to an ECG amplifier, where processing can occur. However, due to the nature of the MRI image acquisition, the patient is subjected to an extreme static magnetic field aligned axially to the patient; moving magnetic gradients in the X, Y, and Z axis; and pulsed radio frequency (RF) fields on the order of 1500V/meter. Each of these fields presents a special challenge for the accurate detection of the “R” wave of the ECG signal.
Considering each of these in order, the static magnetic field in most MRI devices is on the order 1.5 T (tesla) or about 5000 times stronger then the earth's magnetic field. This presents the obvious problem that magnetic items near the bore of the magnet can become projectiles with the resulting patient or clinician injury. There is also a secondary problem with respect to “R” wave detection. Blood, which is conductive, exits the heart orthogonally to the static magnetic field. The movement of blood in the magnetic field results in a “magneto-homodynamic effect”; electrical currents induced in the blood. In other words, because blood is a conductor moving at right angles to the magnetic field, it is equivalent to a generator. The currents induced in the blood will distort the “T” wave, which indicates the re-polarization of the heart, making the “T” wave appear much larger in amplitude than it actually is. This can cause some “R” wave detection algorithms to detect the “T” wave instead, which will cause a shift of some 40 mS from the desired trigger point.
Yet a second problem is caused by the moving magnetic field gradients, since they will cause currents to be generated in any conductor exposed to them. The use of patient leads and cables inside the MRI bore to bring the low level (typically 1 mV) ECG signal to the amplifiers, results in artifacts in the ECG signal that could be in the same bandwidth as the ECG signal itself. The frequency and duration of the gradients are a function of the type of scan sequence being performed, and cannot be filtered effectively with a fixed filter sequence.
Finally, the pulsed RF fields present the greatest challenge to effective ECG detection. The RF pulse is usually a SINC ((Sin x)/x) pulse centered at 64 MHz for a 1.5 T system and is about 5 mS in duration. The repetition rate for the pulses is from tens of Hz to several KHz. The fields are generated within the magnet bore and the coils generating the pulses are excited with 50KW of RF power creating a field strength that often exceeds 1500V/M. These RF pulses, because of the very high power, are a source of significant patient risk when the patient leads are exposed to the fields. A wire, looped around on itself, will appear as a short circuit through the insulator. This in turn will allow for the generation of eddy currents in the loop, which will then heat the wire, often enough to cause third degree burns. To limit the generation of the eddy current, the patient leads must have a distributed impedance of about 10Kohms/ft. The result is that the RF pulses in addition to potentially generating heat in the electrodes create artifacts in the ECG amplifiers at the repetition rate used in a particular scan. The high impedance of the patient leads also increases the electrical noise of the system.
The traditional approach employed by systems presently on the market involve using carbon fiber electrodes attached to high impedance patient leads to bring the low level (1 to 5 mV) ECG signal out of the magnet bore. Once out of the bore, the high impedance leads connect to a conventional patient cable, which then supplies the signal to the ECG amplifiers. The ECG amplifiers typically are located in a RF sealed enclosure outside the magnet and often as far as 15 feet away. Because the signal is heavily contaminated with MRI related artifacts, a huge amount of post processing is required to clean the signal enough to detect the “R” wave. This process is usually DSP (Digital Signal Processing) based, and often requires the use of filters that can change parameters quickly over time to reduce artifacts. These are generally referred to as adaptive filters. Presently, no manufacturer has a solution which can produce a clean ECG waveform under all scan conditions.